The most common chromosomal translocation in liposarcomas, t(12;16)(q13;p11), creates the FUS/TLS-CHOP fusion gene. We previously developed a mouse model of liposarcoma by expressing FUS-CHOP in murine mesenchymal stem cells. In order to understand how FUS-CHOP can initiate liposarcoma, we have now generated transgenic mice expressing altered forms of the FUS-CHOP protein. Transgenic mice expressing high levels of CHOP, which lacks the FUS domain, do not develop any tumor despite its tumorigenicity in vitro and widespread activity of the EF1a promoter. These animals consistently show the accumulation of a glycoprotein material within the terminally di erentiated adipocytes, a characteristic ®gure of liposarcomas associated with FUS-CHOP. On the contrary, transgenic mice expressing the altered form of FUS-CHOP created by the in frame fusion of the FUS domain to the carboxy end of CHOP (CHOP-FUS) developed liposarcomas. No tumors of other tissues were found in these transgenic mice despite widespread activity of the EF1a promoter. The characteristics of the liposarcomas arising in the CHOP-FUS mice were very similar to those previously observed in our FUS-CHOP transgenic mice indicating that the FUS domain is required not only for transformation but also in¯uences the phenotype of the tumor cells. These results provide evidence that the FUS domain of FUS-CHOP plays a speci®c and critical role in the pathogenesis of liposarcoma. Oncogene (2000) 19, 6015 ± 6022.

Fusion proteins created by chromosomal abnormalities are key components of mesenchymal cancer development. The most common chromosomal translocation in liposarcomas, t(12;16)(q13;p11), creates the FUS ± CHOP fusion gene. In the past, we generated FUS ± CHOP and CHOP transgenic mice and have shown that while FUS ± CHOP transgenic develop liposarcomas, mice expressing CHOP, which lacks the FUS domain, display essentially normal white adipose tissue (WAT) development, suggesting that the FUS domain of FUS ± CHOP plays a speci®c and critical role in the pathogenesis of liposarcoma. To test the signi®cance of FUS and CHOP domain interactions within a living mouse, we generated mice expressing the FUS domain and crossed them with CHOP-transgenic mice to generate double-transgenic FUSxCHOP animals. Here we report that expression of the FUS domain restores liposarcoma development in CHOP-transgenic mice. Our results provide genetic evidence that FUS and CHOP domains function in trans for the mutual restoration of liposarcoma. These results identify a new mechanism of tumor-associated fusion genes and might have impact beyond myxoid liposarcoma.