Proto-Oncogenes and Development During the past decade an avalanche of evidence has implicated proto-oncogenes in the control of cell proliferation. Products of proto-oncogenes include many of the secreted growth factors and cell surface receptors that mediate intercellular signaling events (Aaronson, 1991), the components of cytoplasmic pathways that interpret those signals (Cantley et al., 1991), and the nuclear factors that execute their commands through transcriptional regulation (Lewin, 1991). We are now witnessing a similar outpouring of support for the parallel notion that genes important in carcinogenesis are central to pattern formation in the embryo or differentiation in cell lineages. Preexisting or man-made mutations of mouse proto-oncogenes affect a variety of developmental processes (Forrester et al., 1992). Well-established growth factors and oncogenes can induce differentiation under appropriate conditions (eg, Alema et al., 1985). The homologs of proto-oncogenes in flies, worms, and yeast are essential components of developmental mechanisms in those organisms (Hoffmann et al., 1992), and a few genes isolated as determinants of development, such as homeobox-encoding genes, have later been shown to have oncogenic potential (Perkins et al., 1990; Rabbit& 1991). Among the most striking connections between oncogenesis and development are provided by Writ genes, the subject of this review.(The term Writ is an amalgam of wingless [wg] and int [see below and Nusse et al., 19911.) The first Writ gene was cloned from the mouse genome as a relatively obscure proto-oncogene ten years ago (Nusse and Varmus, 1982). But the numerous Writ genes isolated from diverse species in the past few years have now attracted unusual attention because of the dramatic developmental phenomena attributed to them. Ectopic expression of Writ genes induces axis duplication in frog embryos (McMahon and Moon, 1989), as well as mammary cancer in mice; Writ gene deficiencies prevent normal development of mammalian brains (McMahon and Bradley, 1990; Thomas and Capecchi, 1990) and normal segmentation of insect embryos (Rijsewijk et al., 1987a). Any effort to understand these remarkable effects must begin with the fact that Writ genes encode secretory glycoproteins and apparently cause cells to proliferate, to differentiate, or perhaps simply to survive by signaling through autocrine and paracrine routes. In this sense, Wnt proteins are multipotent factors, capable of inducing different biological responses in different cellular contexts, a phenomenon also described for better-known secretory proteins such as the fibroblast growth factors (FGFs), the transforming growth factors p (TGFB), and nerve growth factor (NGF). Establishing a more detailed picture of the actions of Wntgenes, however, presents considerablechallenges. The Writ gene family is large, suggesting functional redundancy, the patterns of expression are complex, proteins encoded by Writ genes have resisted isolation in biologically active form, and receptors for Wnt proteins, presumably essential components in cell-cell signaling, remain to be identified.

The Writ Gene Family With the benefit of hindsight, we now recognize that phenomena studied for several decades are the consequences of Writ gene mutations. Viral insertion mutations regularly promote mammary tumors in laboratory mice (Bittner, 1938; Korteweg, 1936), a spontaneous frameshift mutation of mice (swaying) impairs cerebellar structure and function (Lane, 1967; Thomas et al., 1991), and wg mutations in Drosophilacan transform a wing to a notum or disrupt segment polarity (Sharma, 1973 …