The Fus1 gene resides in the critical 3p21.3 human chromosomal region deleted in lung and breast cancers. Recently, the tumour suppressor properties of Fus1 were confirmed experimentally by intra-tumoural administration of Fus1 that suppressed experimental lung metastasis in mice. We generated Fus1-deficient mice that were viable, fertile, and demonstrated a complex immunological phenotype. Animals with a disrupted Fus1 gene developed signs of autoimmune disease, such as vasculitis, glomerulonephritis, anaemia, circulating autoantibodies, and showed an increased frequency of spontaneous vascular tumours. Preliminary analysis of immune cell populations revealed a consistent defect in NK cell maturation in Fus1 null mice that correlated with changes in the expression of IL-15. Injection of IL-15 into Fus1 knockout mice completely rescued the NK cell maturation defect. Based on these results, we propose the hypothesis that Fus1 deficiency affects NK cell maturation through the reduction of IL-15 production but does not directly alter their developmental capacity. Since acquired immunity was not affected in Fus1-deficient animals, we suggest a relationship between the Fus1 protein and the regulation of innate immunity via IL-15 production. The increased frequency of spontaneous cancers and the development of an autoimmune syndrome in Fus1 null mice imply that these mice could serve as a model for studying molecular mechanisms of anti-tumour immunity and autoimmunity.

The murine Ly49 gene family encoding natural killer cell receptors for class I MHC is an example of a rapidly evolving cluster of immune response genes. Determining the genomic sequence of the 129S6/SvEvTac (129S6) Ly49 cluster and comparing it to the known sequence of the C57BL/6 (B6) region provided insight into the mechanisms of Ly49 gene evolution. 129S6 contains 20 Ly49, many of which are pseudogenes and 40% of the genes have no counterpart in the B6 genome. The difference in gene content between these two strains is primarily the result of distinct patterns of gene duplication. Phylogenetic analyses of individual exons showed that Ly49 genes form distinct sub-families and an ancestral haplotype can be surmised. Dotplot analysis supports limited allelism in the two haplotypes; however, large regions of variation punctuate these islands of colinearity. These variable regions contain a high concentration of repetitive elements that are predicted to contribute to the dynamic evolution of this cluster. The extreme variation in Ly49 haplotype content between mouse strains provides a genetic explanation for the documented differences in natural killer cell phenotype, and also indicates that differences in natural killer cell function observed between B6 and 129-derived gene-targeted mice should be interpreted with caution.