Cathepsin-D, a lysosomal aspartyl proteinase, is highly secreted by breast cancer cells and its over-expression by transfection stimulates cancer cell proliferation. The mechanism by which this protease a ects proliferation remains, however, unknown. In order to determine whether proteolytic activity is necessary, we abolished its enzymatic activity using site-directed mutagenesis followed by stable transfection in 3Y1-Ad12 cancer cells. Substitution of the aspartic acid residue 231 by an asparagine residue in its catalytic site abrogated the cathepsin-D proteolytic activity but did not a ect its expression level, processing or secretion. However, like wild-type cathepsin-D, this mutated catalytically-inactive cathepsin-D retained its capacity to stimulate proliferation of cells embedded in Matrigel or collagen I matrices, colony formation in soft agar and tumor growth in athymic nude mice. Addition on the mocktransfected cells, of either conditioned media containing the wild-type or the mutated pro-cathepsin-D, or of the puri®ed mutated pro-cathepsin-D, partially mimicked the mitogenic activity of the transfected cathepsin-D, indicating a role of the secreted pro-enzyme. Moreover, addition of two anti-cathepsin-D antibodies on the cathepsin-D transfected cells inhibited their proliferation, suggesting an action of the secreted pro-cathepsin-D via an autocrine loop. A synthetic peptide containing the 27-44 residue moiety of the cathepsin-D pro-fragment was, however, not mitogenic suggesting that a receptor for the pro-fragment was not involved. Furthermore, the cathepsin-D mitogenicity was not blocked by inhibiting the interaction of pro-cathepsin-D with the mannose-6-phosphate receptors. Our results altogether demonstrate that a mutated cathepsin-D devoid of catalytic activity is still mitogenic and suggest that it is acting extracellularly by triggering directly or indirectly a yet unidenti®ed cell surface receptor. Oncogene (2001) 20, 6920 ± 6929.