In patients, chronic treatment with lithium leads to renal microcysts and nephrogenic diabetes insipidus (NDI). It was hypothesized that renal cyclooxygenase-2 (COX-2) activity promotes microcyst formation and NDI. Kidney microcysts were induced in male adolescent rats by feeding dams with lithium (50 mmol/kg chow) from postnatal days 7-34. Lithium treatment induced somatic growth retardation, renal microcysts and dilatations in cortical collecting duct; it increased cortical cell proliferation and inactive pGSK-3β abundance; it lowered aquaporin-2 (AQP2) protein abundance and induced polyuria with decreased ability to concentrate the urine; and it increased COX-2 protein level in thick ascending limb. Concomitant treatment with lithium and a specific COX-2 inhibitor, parecoxib (5 mg/kg per day, P10-P34), did not prevent lithium-induced microcysts and polyuria, but improved urine concentrating ability transiently after a 1-desamino-8-D-arginine vasopressin challenge. COX-2 inhibition did not reduce cortical lithium-induced cell proliferation and phosphorylation of glycogen synthase kinase-3β (GSK-3β). COX-1 protein abundance increased in rat kidney cortex in response to lithium. COX-1 immunoreactivity was found in microcyst epithelium in rat kidney. A human nephrectomy specimen from a patient treated for 28 years with lithium displayed multiple, COX-1-immunopositive, microcysts. In chronic lithium-treated adolescent rats, COX-2 is not colocalized with microcystic epithelium, mitotic activity, and inactive pGSK-3β in collecting duct; a blocker of COX-2 does not prevent cell proliferation, cyst formation, or GSK-3β inactivation. It is concluded that COX-2 activity is not the primary cause for microcysts and polyuria in a NaCl-substituted rat model of lithium nephropathy. COX-1 is a relevant candidate to affect the injured epithelium.