The sphingolipid activator proteins (saposins A, B, C and D) are small homologous glycoproteins that are encoded by a single gene in tandem within a large precursor protein (prosaposin) and are required for in vivo degradation of some sphingolipids with relatively short carbohydrate chains. Human patients with prosaposin or specific saposin B or C deficiency are known, and prosaposin- and saposin A-deficient mouse lines have been generated. Experimental evidence suggests that saposin D may be a lysosomal acid ceramidase activator. However, no specific saposin D deficiency state is known in any mammalian species. We have generated a specific saposin D(-/-) mouse by introducing a mutation (C509S) into the saposin D domain of the mouse prosaposin gene. Saposin D(-/-) mice developed progressive polyuria at around 2 months and ataxia at around 4 months. Pathologically, the kidney of saposin D(-/-) mice showed renal tubular degeneration and eventual hydronephrosis. In the nervous system, progressive and selective loss of the cerebellar Purkinje cells in a striped pattern was conspicuous, and almost all Purkinje cells disappeared by 12 months. Biochemically, ceramides, particularly those containing hydroxy fatty acids accumulated in the kidney and the brain, most prominently in the cerebellum. These results not only indicate the role of saposin D in in vivo ceramide metabolism, but also suggest possible cytotoxicity of ceramide underlying the cerebellar Purkinje cell and renal tubular cell degeneration.