Search

In the renal collecting duct, vasopressin acutely activates cAMP production, resulting in trafficking of aquaporin-2 water channels (AQP2) to the apical plasma membrane, thereby increasing water permeability. This acute response is modulated by long-term changes in AQP2 expression. Recently, a cAMP-responsive element has been identified in the AQP2 gene, raising the possibility that changes in cAMP levels may control AQP2 expression. To investigate this possibility, we determined AQP2 protein levels in a strain of mice, DI +/+ severe (DI), which have genetically high levels of cAMP-phosphodiesterase activity, and hence low cellular cAMP levels, and severe polyuria. Semiquantitative immunoblotting of membrane fractions prepared from whole kidneys revealed that AQP2 levels in DI mice were only 26 +/- 7% (+/-SE) of those in control mice (n = 10, P < 0.01). In addition, semiquantitative Northern blotting revealed a significantly lower AQP2 mRNA expression in kidneys from DI mice compared with control mice (43 +/- 6% vs. 100 +/- 10%; n = 6 in each group, P < 0.05). AQP3 levels were also reduced. The mice were polyuric and urine osmolalities were accordingly substantially lower in the DI mice than in controls (496 +/- 53 vs. 1,696 +/- 105 mosmol/kgH2O, respectively). Moreover, there was a linear correlation between urine osmolalities and AQP2 levels (P < 0.05). Immunoelectron microscopy confirmed the markedly lower expression of AQP2 in collecting duct principal cells in kidneys of DI mice and, furthermore, demonstrated that AQP2 was almost completely absent from the apical plasma membrane. Thus expression of AQP2 and AQP2 trafficking were severely impaired in DI mice. These results are consistent with the view that in vivo regulation of AQP2 expression by vasopressin is mediated by cAMP.

One of the mechanisms by which Li evokes polyuria is thought to be impairment of arginine vasopressin (AVP)-sensitive adenylate cyclase (AdC) in cells of the renal collecting duct. To investigate how AdC is influenced by chronic administration of Li, we created nephrogenic diabetes insipidus (NDI) in rats and microdissected the medullary collecting tubule from both control and NDI rats. In the NDI group, the 10(-6) M AVP-stimulated cAMP contents failed to increase completely, and the levels were significantly lower than that of the control group (10.4 +/- 1.4 vs. 48.4 +/- 4.7 fmol/mm, P less than 0.001). Pretreatment with pertussis toxin (PT), an inhibitor of inhibitory G protein (Gi), did not affect the basal cAMP levels in both groups, although it increased AVP-stimulated cAMP production in the NDI group in a dose- and time-dependent manner. AVP-stimulated cAMP production with over 100 ng/ml PT in the NDI group reached the levels observed in the control group. Incubation with cholera toxin, an agonist of stimulatory G protein (Gs), increased the cAMP content in the two groups to almost equal levels. To exclude the possibility that prostaglandin E2 (PGE2) is involved in the cellular mechanism of Li-induced NDI, the effect of indomethacin (Indo) on PT action was examined. However, Indo (10(-5) M) did not influence either the basal or AVP-dependent cAMP contents. From these results it is suggested that Li impairs AVP-sensitive AdC not through inhibition of Gs but through activation of Gi and that PGE2 may not be involved in the cellular pathogenesis of NDI at least in the rat at the step of cAMP formation.