To investigate the expression of aquaporins-3 (AQP3) in amniotic epithelial cells regulated by cyclic adenosine monophosphate-protein kinase A (cAMP-PKA) signal pathway and to explore the mechanisms of its expression.The amniotic epithelial cells were collected from 30 patients who underwent elective caesarean sections at term with normal amniotic fluid volume and primarily cultured. The cultured cells were treated with (1)forskolin groups: different concentration (0, 2.5, 5, 50 or 100 µmol/L) of forskolin treated cells for 2 hours, and the optimal concentration of forskolin treated cells with different time (0, 1, 2, 10 or 20 hours); (2)SP-cAMP groups: different concentration (0, 2.5, 5, 50 or 100 µmol/L) of SP-cAMP treated cells for 2 hours, and the optimal concentration of SP-cAMP treated cells with different time (0, 1, 2, 10 or 20 hours); (3)H-89 groups: different concentration (0, 5, 10, 50 or 100 µmol/L) of H-89 treated cells for 2 hours, and the optimal concentration of H-89 treated cells with different time (0, 1, 2, 10 or 20 hours ). The level of intracellular cAMP and activity of PKA were detected by using ELISA, and immunohistochemistry was used to detect the localization of AQP3, the protein expression of total cAMP-response element binding protein (CREB) and phospho-CREB (p-CREB) and AQP3 were assessed by western blot analysis. Cell proliferation was assessed by cell counting kit-8 (CCK-8) assay.(1) The brown staining of AQP3 was detected in both cell membrane and cytoplasm in each group. (2) There was no significant change of the cell proliferation rate among groups with different concentration of forskolin, SP-cAMP and H-89 treatment (P0.05). (3) After different concentration of forskolin treated 2 hours, the expression of total CREB had no significant difference among them(P0.05). While the expression of cAMP level, PKA activity, p-CREB and AQP3 protein were significantly changed, which were higher in 2.5 µmol/L, 5 µmol/L, 50 µmol/L forskolin group when compared with 0 µmol/L (P0.05). Their expressions in 5 µmol/L forskolin group were higher than that in 2.5 µmol/L and 50 µmol/L (P0.05). The optimal forskolin concentration was 5 µmol/L. (4) After different concentration of SP-cAMP treated 2 hours, the expression of total CREB and cAMP level had no significant difference among them (P0.05), while the expression of PKA activity, p-CREB and AQP3 protein were significantly changed, which were higher in 5 µmol/L, 50 µmol/L SP-cAMP group when compared with 0 µmol/L (P0.05). Their expressions in 50 µmol/L SP-cAMP group were higher than that in 5 µmol/L (P0.05). The optimal SP-cAMP concentration was 50 µmol/L. (5) After different concentration of H-89 treated 2 hours, the expression of total CREB and cAMP level had no significant difference among them (P0.05), while the expression of PKA activity, p-CREB and AQP3 protein were significantly changed, which were lower in 10 µmol/L, 50 µmol/L and 100 µmol/L H-89 group when compared with 0 µmol/L (P0.05). Their expressions in 10 µmol/L H-89 group were lower than that in 50 µmol/L, 100 µmol/L (P0.05). The optimal H-89 concentration was 10 µmol/L. (6) p-CREB and AQP3 protein expression were significantly lower in 5 µmol/L forskolin combined 10 µmol/L H-89 incubating 2 hours group when compared with 5 µmol/L forskolin, but higher than that in 10 µmol/L H-89 treated group (P0.05). Total CREB was no significant difference among the three groups (P0.05).cAMP-PKA signal transduction pathway may regulate AQP3 protein expression in human amniotic epithelial cells.